Back to the feedChatGPT Promo Codes
Ends Feb 28

Ultrafast list price is 6x Standard: Astra $48 input / $240 output per million tokens

OpenAI GPT-6.2 Astra and Sol Ultrafast API at half price

Ultrafast mode for GPT-6.2 Astra and Sol at half price, billed at 3x the Standard rate

OpenAI has cut Ultrafast pricing for GPT-6.2 Astra and Sol in half. API and Enterprise users pay 3x the Standard rate instead of 6x through February 28, 2027 (Pacific time).

Eligible plans
OpenAI API pay-as-you-go and Enterprise
Eligibility
OpenAI API and Enterprise users; the discounted rate applies automatically
Benefit duration
Through February 28, 2027 (Pacific time)
Renewal terms
From March 1, 2027, Ultrafast returns to 6x the Standard rate
Claim by
2027-02-28

The briefing

OpenAI has released GPT-6.2 Astra and GPT-6.2 Sol, and both support the Fast and Ultrafast speed modes. Alongside the launch, OpenAI cut Ultrafast API pricing in half for the next three months, so it now costs 3x the Standard rate.

How much you save

Ultrafast normally costs 6x the Standard rate. Through February 28, 2027 (Pacific time) it is billed at 3x, and it returns to 6x on March 1, 2027. Prices per million tokens for requests up to 272K input tokens:

  • GPT-6.2 Astra drops from $48 to $24 for input, from $240 to $120 for output, and from $2.40 to $1.20 for cached input.
  • GPT-6.2 Sol drops from $12 to $6 for input, from $60 to $30 for output, and from $0.60 to $0.30 for cached input.

Long-context requests above 272K tokens are also billed at half the Ultrafast list price.

Who gets it

OpenAI API users and Enterprise customers get the discounted rate automatically, with nothing to claim. On the ChatGPT side, Pro 500 subscribers can use Ultrafast with both models, and Pro 200 subscribers get it first with GPT-6.2 Sol.

How to use it

  1. Set the model to gpt-6.2-astra or gpt-6.2-sol in your API request.
  2. Turn on Ultrafast mode as described in the API docs, and usage is billed at the promotional rate.

See the OpenAI API pricing page for all rates and the GPT-6.2 Astra and Sol announcement for model details.

Is this still live?

Community signal100% say it still works